|
Activity Number:
|
476
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #305049 |
|
Title:
|
Analysis of Method Comparison Studies with Multiple Methods
|
|
Author(s):
|
Pankaj K. Choudhary*+ and Kunshan Yin
|
|
Companies:
|
The University of Texas at Dallas and EQECAT
|
|
Address:
|
Department of Mathematical Sciences, EC 35, Richardson, TX, 75080,
|
|
Keywords:
|
Agreement ; Concordance correlation ; Mixed effects model ; Multiple Comparisons ; Tolerance interval ; Total deviation index
|
|
Abstract:
|
Evaluation of agreement among multiple methods of clinical measurement is a topic of considerable interest in health sciences. As in an analysis of variance comparing more than two treatment means, when more than two measurement methods are compared, it is of primary interest to perform multiple comparisons and rank pairs of methods on the basis of their extent of agreement. We develop frequentist and Bayesian procedures for this purpose. In particular, simultaneous confidence and credible bounds are developed for multiple comparisons, and two complementary approaches are developed for ranking method pairs -- one based on simultaneous bounds and the other based on posterior probabilities. The procedures can be used with any scalar measure of agreement. Their small-sample performance is evaluated using simulation. A real example from the literature is provided for illustration.
|
